2016 范德博 Marinus VAN DEN BERG

Home , Beijing dialect, Hainanese, Shanghainese, Taishanese

Proceedings of the 29th North American Conference on Chinese Linguistics (NACCL-29). 2017. Volume 1. Edited by Lan Zhang. University of Memphis, Memphis, TN. Pages 85-98.

THE CHINA URBAN LANGUAGE SURVEY PROJECT 2003 - 2016

范德博 Marinus VAN DEN BERG Leiden University, IIAS, Emeritus

In this presentation, three major hypotheses developed in the context of the China Urban Language Survey Project will be discussed. Changes in the urban language environment are mainly related to large scale migration from the country side and other places toward the newly developed and developing industrial centers in China’s southeastern provinces. The project started in 2003 in Nanjing University’s Sociolinguistic Laboratory under the guidance of professor Xu Daming, attracted researchers from various universities both in China, Hong Kong, Taiwan and Europe, and obtained funding from various sources including a major grant from the Netherlands’ Organization for Scientific Research (NWO). Progress reports were presented in yearly conferences, and major findings appeared in a 2010 book publication (van den Berg & Xu 2010). We will start this presentation with the notion “long-term accommodation”, based on work by Van den Berg in Taiwan in 1977-78 (van den Berg 1986), extend that notion to developments in mainland China, and present three major hypotheses developed in the context of the China Language Survey Project, a triglossia/diglossia hypothesis based on work in Hainan province (Tsou et al. 2010), the fundamentals of Speech Community Theory developed in studies of the language situation in the Inner-Mongolian city of Baotou (Xu 2004; 2010), and the concept of network density, developed in a study of the language situation in Beijing (Song & Zhu 2016).

1. Long-term accommodation In a questionnaire survey among National Taiwan University students in 1977, it was possible to demonstrate across generation adjustment to the national language configuration at each of the time frames involved. For grandparents of the students, who were born during the Qing Dynasty or during the beginning years of the Japanese colonial period, Minnan was the dominant language, and more so for grandmothers (71%; 74%) than for grandfathers (52%; 48%). The first of these figures giving maternal grandparents language backgrounds, and the second those of students’ paternal grandparents. These data confirm the limited options for female education during the last of the empirical dynasties, whereas Minnan-Japanese bilingualism data show the first signs of that gender gap’s closure. For maternal grandparents, the bilingualism data were 22% and 27% for

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

grandmothers and grandfathers respectively, whereas Minnan-Japanese data for paternal grandparents showed a wider gap, 21% for the grandmothers, comparable to the 22% found for the maternal grandmothers, and 35% for the grandfathers, extensively higher than the figure found for the maternal grandfathers (27%). The interesting observation for the generation of parents born during the Japanese period is that a majority of them became trilingual in Minnan, Japanese, and Mandarin Chinese, called Guoyu 國語 under the Republican government. A gender gap still existed, but both sexes had in majority become trilingual. The data for the students’ mothers was 46% and for the fathers 57%. Parents who were older or were otherwise disadvantaged in learning a new language variety did not pick-up Guoyu 國語 , and remained bilingual Minnan-Japanese speakers, and this occurred more typically for mothers (18%) than for fathers (12%). Minnan mono-lingualism showed a similar gender difference, 20% for mothers and 9% for fathers. These latter differences most likely have a correlation with economic activity (farm labor) and income. The language situation of the students themselves gives a totally different picture. Japanese has disappeared from the language repertoire, and the best language claimed by almost all students (93%) is Mandarin Chinese (Guoyu). With an average age of 20, these students were born around 1957, and had participated in a Mandarin Chinese dominant education system, moving from elementary school, to high school, and on to university, in this case, one of the top universities in Taiwan, implying that the results are not those of the average student, but of a selection of Taiwan’s top students, who also studied in Taiwan’s political, economic, and educational center, Taipei. Keeping this in mind, these data allow the interpretation that members of a national community adjust to the national level language configuration of that community. This adjustment, using insights from interpersonal accommodation theory (Giles & Powesland 1986), we called long-term accommodation, suggesting that over-time community members, while keeping accommodating to everyday language requirements, are forced to adjust to the norms set at the national level. The result is not unexpected. It was observed for the development of Latin in the Roman Empire and related to factors such as a central government supporting economic development, making it worthwhile to acquire the language of government and education, the presence of social mobility, which helps to create multilingual areas, thereby creating the need for a lingua franca. Using this insight and turning now to mainland China, it is possible to predict that over- time the national standard language, Putonghua, will spread. The conditions for that spread, using the Taiwan and Roman data, are a well-organized education system, a language market supporting the use of Putonghua, economic development, and social mobility which will make Putonghua the aspired to lingua franca. As we will see, education got reorganized after 1980, the language market was strongly influenced by mass-migration, and the spread of Putonghua awaited the emergence of an economically developing China.

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

2. Developments in Modern China Without paying attention to the pre-1980 language situation, we must conclude that the long-term accommodation hypothesis is not confirmed by developments in the 1980s. The language market in the first ten years after the start of the 1980 Open Door policy was in favor of Cantonese, the language variety of the capitalist entrepreneurs with a Hong Kong background, who invested in factories in Shenzhen and the Pearl River Delta (Zhan 1993; Guo 2004). The first wave of workers came from the surrounding country-side and were all speakers of some Cantonese dialect, and as a result oriented themselves on Cantonese rather than on Putonghua for communication. Shenzhen, the first Special Economic Zone, and the surrounding areas soon also attracted many people from other parts of China, including the Mandarin dialect area. Migrants in a new city like Shenzhen mixed home dialect languages (Hakka, Siyi (Taishanese), Swatow (Shantou), etc.) with Putonghua and Cantonese, making Putonghua the dominant language for the technical professions and for business transactions in new districts (Van den Berg 2009; Tang 2016), whereas in Cantonese speaking Guangzhou, northern and better educated migrants formed their own Putonghua based networks (Van den Berg 2010). One question that in this setting comes to the fore is how will these developments work-out for China as a whole? Various answers are possible, Putonghua will destroy the regional dialects, the dialects will stop the spread of Putonghua, or a new bilingual balance will develop. We will not discuss each of these possibilities further at this moment, but in the following we will discuss three proposals, the triglossia/diglossia hypothesis, Speech Community Theory, and network density. Each gives a somewhat different view on what is most likely the future development of the Chinese urban language market. The first of these is the strongest proposal and is the one that addresses the national language situation. It predicts a general tendency, so let’s see what that line of research has to say.

3. From triglossia to diglossia In an elementary school survey in Sanya, Hainan province, researchers observed that students, in addition to some use of the regional language variety Hainanese, mainly used the home dialect (six different language varieties), when talking to their grandparents, whereas in communication with parents, some Mandarin Chinese was introduced. When talking to each other the level of Mandarin Chinese got more extensive, reaching even higher levels when an everyday task such as shopping was involved. In public transportation, finally, Mandarin Chinese obtained its maximum use. In the latter case, we can imagine the lingua franca effect of the use of Mandarin Chinese, particularly when public transportation personnel have different backgrounds. Hainanese is still being used as lingua franca for around 20 percent of the cases, but in the remaining settings Putonghua is dominant, taking over in effect the lingua franca role of Hainanese (Tsou et al. 2010). The researchers concluded that given rapid modernization taking place in Sanya, there is a language shift taking place from the home language to, what they call, the

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

Supreme Language, Putonghua. They see this as a shift from a triglossia situation (home dialect-Hainanese-Putonghua) toward a diglossia setting (home language-Putonghua). And they even went one step further claiming that this development is taking place throughout China, the motivation being that a triglossic setting demands a higher psychological burden than a diglossic setting, and as a result it would be only natural to see a shift in the direction of diglossia (home dialect-Supreme language), whereas in the long run the home dialects too will have to disappear, mono-lingualism assumedly providing even less of a cognitive burden. Given this analysis, there are quite a few questions that can be asked, and we will do so in a moment, when we will address the language situation in Shanghai. One element, however, that from the perspective of urban language studies is clearly missing in the discussion is the impact of social stratification, and in order to get that more clearly into perspective it is necessary to focus on Speech Community Theory, which states that in an economically developing urban environment social stratification and language differences must occur.

4. Speech Community Theory Bloomfield, as early as 1933, devoted a full chapter to the discussion of the speech community (Bloomfield 1933: 42-56). He defined a speech community as: “a group of people who interact by means of speech” (Bloomfield 1933: 42). Simple as this definition might seem, Bloomfield was clearly aware of many of its implications, for instance, he did not specify what kind of speech, and in this definition, this can be any kind of speech, the fundamental and essential ability of humans to interact with each other. He further specified the speech community “as the most important kind of social group”, different from “other phases of social cohesion, such as economic, political or cultural groups”. More importantly, quite early in the chapter, Bloomfield mentioned “the assimilation into a speech community of whole groups of foreigners, such as immigrants, conquered people, or captives,” an issue that is at the heart of the present presentation. Unfortunately, he has not much to say about the process itself, but he is clearly aware of the implications. Limiting ourselves to complex, large sized, speech communities, a distinction Bloomfield makes too, he observes that everywhere differences in speech develop, the main sources are distance, which is geographical, and social differentiation. As to the latter, he points to the speech differences which develop within the standard language under the influence of differences in family tradition, schooling, occupation, and income, which results in subgroups we recognize as social classes. Of course, in this general introduction, Bloomfield did not discuss details of the formation process of a speech community. That issue was taken up more recently in a series of real-time studies I like to introduce now, and which we know under the title of Speech Community Theory (Xu 2004; 2010; 2016). The observations by Bloomfield got a new impetus when the New York speech community was analyzed in an empirical way (Labov 1966/2006). That approach

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

confirmed in essence most of the distinctions made by Bloomfield, but now certain of these distinctions were given real content such as a precise distribution of the phonemes [-r] and [aw] across social space. Often too new labels, such as style shifting, were introduced. This empirical approach became the new standard, and the empirical study of a speech community was fundamentally different from earlier more theoretically oriented approaches (Patrick 2002). It is this line of research that was taken up in the study of a developing Chinese speech community, the new Kundulun district in the city of Baotou in Inner Mongolia (Xu 2010). Speech community theory starts from the observation that the social system of speakers (think of the social groups mentioned by Bloomfield) is the basis for understanding the way of speaking in a community. Where linguists tend to focus on speech sounds and grammatical rules, they lose contact with the social organization of the speaker group. In order to clarify this, Xu Daming and his students over a period of twenty years studied the changing language situation in the already mentioned new industrial district of Kundulun, part of the city of Baotou in Inner- Mongolia. This district was established in 1956 as the residential area for workers of the large state-owned Baotou Steel Corporation, housing around three-hundred thousand people. Migrants came from all parts of China, but the majority originated from surrounding provinces, and spoke varieties of Mandarin dialects. Dialect contact resulted in variation in forms of nasalization. A comparison of the results of two studies, one in 1987 and a follow-up study in 2006, showed that the relationship between internal phonetic variation and selected social variables gets more complex over time, and this led to the formulation of a theory of speech community formation. The theory stresses the development of shared forms of communication as the result of day-to-day communication between same-group community members, compare Bloomfield’s mutual adjustment. The relation between ‘occupation’ and other variables (‘social network’, ‘place of origin’ as the more persistent ones) with phonetic variation made clear that, since occupation, education, and income are correlated, this kind of variation creates a socially stratified society. And in such a society, speech styles are stratified too. The importance of these studies is that they trace, in great detail, the development of a new speech community in terms of intrinsic constraints in relation to a set of social variables. The implication for the triglossia>diglossia hypothesis is that is needs to be evaluated in the context of a socially stratified society, and that is what we will do in the next section, where we will study a large, complex, socially stratified, speech community and address the issue of migration as raised by Bloomfield and see to what extent this leads to assimilation, as he noted, or will take different forms. The analysis will also allow us to test to what extent the triglossia hypothesis put forward in the previous section can be maintained.

5. The Shanghai language situation Neglecting the formative period of the Shanghai dialect in the 19th and 20th centuries (see for this issue Van den Berg 2016), the developing Shanghai language

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

situation after the civil war, can be divided into two periods, a relatively stable period after 1949 and a rapidly changing period with massive migration after 1990. Let us start by looking at the demographic data in order to see when and to what extent migration occurred. The population development data since 1952, shown in figure 1, shows a curve that at first during the beginning years of communist control mildly increased, then, during the cultural revolution, fell back under the influence of the xia xiang, ‘back to the country side’ policy. After 1990, the curve rises rapidly, increasing from 12 million at that time to 23 million in 2010, an increase of almost hundred percent when compared with 1980, the year the ‘Open Door’ policy started. In the first period after 1949, Shanghai was economically strangled by the new power holders, as a punishment for its days of glory and Western involvement in the previous hundred or so years. That first period was the time of social control and the ‘iron rice bowl,’ and social stratification the result of family tradition, occupation, schooling and income did not shape Shanghai society, unless we accept a class difference between communist cadres and the population at large. In any way, that would be a completely different form of stratification from what we saw developing after 1990, when Shanghai was given the opportunity to develop again.

Population development Shanghai Shanghai pop from 1950; 5 y intervals

35000000

30000000

25000000

Shanghai Year 20000000 Shanghai Populatiom

15000000

10000000

5000000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Figure 1. Shanghai’s population development after 1950

After 1990, Shanghai started to develop again after it finally was allowed to do so, and sprinted forward, becoming China’s economic center again in just a few years. In this

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

second period, international cooperation intensified, the stock market reopened, new occupations were added, and higher incomes obtained. This happened in particular after 2000 when university education had modernized, and the younger generation could compete in many fields. The result was a stratified society, with the under forty generation developing most rapidly. Language use data in four economically stratified department stores in the Xujiahui area support that view (see Van den Berg 2016 for the details). In the second period, one of the driving forces in Shanghai’s rise to prominence was migration. The 2010 census showed that of Shanghai’s 23 million inhabitants, 9 million were migrants with long-term resident status, and those with origins from the Mandarin dialect areas Anhui (29%), Jiangsu (17%), Henan (9%), and Sichuan (7%) were the largest groups. The 2010 census also showed that four-fifth of the migrants had a rural background, suggesting that at that time manual labor was still needed most, but that higher educated groups were increasing in number. Data of language use in public places shows that the manual labor group communicated in their home dialects or used an accented form of Mandarin Chinese. It is the second group that in its various daily contacts, both inside the job, and during shopping or leisure, uses Putonghua, the language of higher education (van den Berg 2016). This form of communication is possible only when, at the Shanghainese side, bilingualism has developed as well. Assuming that is the case, we can map the various social groups and their level of bilingualism as four social classes (Figure 2).

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

Language Stratification in Shanghai Statistics copied from New York 1963 (Labov, 1963; table 7.5)

Figure 2. Social class and language stratification in Shanghai

At the top we have the one percent elite group, representing very rich entrepreneurs, in majority speaking Putonghua, but also people speaking their home dialect, Shanghainese, a Wu dialect, or a Mandarin (Guanhua) dialect. The majority of the population (50%), however, belongs to the lower and upper working classes, and it is they who speak in majority Shanghainese, or one of their many home dialects, be it from the Wu group or the Mandarin group. Among them, and most likely more extensively among the upper working class, various people have a high school background and are accustomed to speaking Putonghua. The next largest group is the lower middle class (40%), education among them is higher and so is their income, their level of bilingualism and the use of Putonghua. In this group, as the figure suggests, better educated Shanghainese speakers, Wu dialect and Mandarin dialect speakers, tend to match the group of Putonghua speakers. Our fourth and last group (9%) represents members of the upper-middle class, these are all higher educated individuals, the majority of which is accustomed to speaking Putonghua. Upper-middle class membership, however, is not restricted to Putonghua speakers, we also find people with a Shanghainese background, who prefer to speak Shanghainese. Most Wu dialect speakers will switch to Shanghainese when interacting with locals, switching to Putonghua when appropriate. In the same way, Mandarin dialect speakers, in an interaction with a local person, most likely will use Putonghua, helping to increase that percentage that way. When talking to a guxiang, ‘home dialect person’ the shared dialect will be used (Fig. 2). Having established the status of social diversification in modern Shanghai society, we now need to determine the extent to which the triglossia>diglossia hypothesis holds in this context? Shanghainese and some of the Wu and Mandarin dialects are well being

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

maintained among members of the working class, either by themselves or with support from the home county or township. The chances that Shanghainese, as the regional language, and the one comparable to Hainanese, will be given up, however, is small. Shanghainese is a high-status language variety and will be maintained for the foreseeable future by millions of people. Wu dialects spoken in Shanghai are under pressure from this more prestigious Shanghainese, whereas Mandarin dialects will feel the pressure from the Supreme variety, Putonghua. Both dialect groups, Wu dialects and Mandarin dialects, however, have strong backing from their home dialect areas and those areas will not be directly influenced by what happens in Shanghai. These various home dialects might be used less in Shanghai but will remain supported at the home county area. They remain under influence of national level education policy though. The triglossia>diglossia hypothesis therefore is not supported by the Shanghai data, and a reformulation will be necessary. What is needed, in addition to the urban phenomenon of social stratification, -- the Hainan survey does no address that issue adequately either -- is a distinction which describes differences in vitality of the language varieties in use in Shanghai. A language variety’s vitality is determined by three factors, status factors, demographic factors, and institutional support factors (Bourhis et al., 1981). Shanghai’s status is related to Shanghai’s position as China’s main economic center. In Hainan that would be a comparison with Haikou, the Hainan provincial capital, but that information was not provided in the Tsou et al. study. Demographic data further set Shanghainese apart from Hainanese, which attracted far smaller numbers of speakers. The third factor, institutional support, favors Putonghua as the Supreme variety, with support in both education and the media. Shanghainese, however, still has a certain amount of support in those domains, whereas that support might be there for Hainanese, but was not documented in the quoted paper and we assume that support is limited. Given this observation, what we see in Shanghai is the demise of home language varieties, the ones that are passed on from generation to generation, since it is the modern younger generation who are modernizing rapidly and come to see those varieties as superfluous. In Shanghai it is not the regional language that is pushed out under influence of Putonghua, Shanghainese has high status and is maintained, what did change was the level of bilingualism among Shanghainese speakers, which strongly increased. What does tend to disappear, in contrast, is the original home dialects in Shanghai (Chu 2001). These dialects have little vitality, in terms of status, number of speakers, and institutional support, and are under pressure from both the regional language Shanghainese and the Supreme language Putonghua. It is in particular the younger generation that in this rapidly changing modern society find it difficult to support these home language varieties, despite their emotional link to the home county, especially when ancestor worship still is located there. Having set the stage for bilingualism in a multi-million and multi-lingual city as Shanghai, it is now time to look at bilingualism in another multi-million city, Beijing.

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

How will networks function in that city, and what can we learn from it for the triglossia>diglossia hypothesis and for social stratification research?

6. Network density Bloomfield observed that in a complex speech community we can observe differences in density of communication. Some speakers communicate more to certain contacts than to others. Theoretically, these differences could be mapped, even patterns across-time could be discovered, but, as Bloomfield observed, in reality this is impossible to do (1933: 42-56). Taking up this challenge, Song and Zhu (2016) designed a technique for comparing the density of communication among young (20-44 years) native Beijing dialect speakers, with the intent to being able to evaluate the strength of maintenance of the Beijing dialect versus the incoming force of Modern Standard Chinese, Putonghua. Using the network concept, they called a person’s density of communication with other contacts, his network density. Their procedure was to select a sample of Beijing dialect speakers (n=269) from each of the ten central city districts, five inner-city districts and five inner- suburb districts, and asked respondents to select five persons with whom they communicated most frequently on a daily basis. They also asked the respondents to order these contacts according to the amount of time they were communicating. This procedure resulted in a classification that compared the number of Beijing dialect speaking contacts in a respondent’s network of five, with the number of Putonghua speaking contacts in that same network. Using the abbreviation of BAF for Beijing Accent Friend, and PAF for Putonghua Accent Friend, this classification matched a BAF of 5 (all five contacts speaking Beijing dialect) with a PAF score of 0 (zero), and a BAF score of 4 with a PAF score of 1, etc. Respondents data were collected either in face-to-face encounters or through telephone interviews. The procedure chosen was time-consuming but rewarding. A standard set of questionnaire questions was used to encourage the respondents to comment on their age and background, how long they were living in that particular district, detail their attitude toward Beijing dialect and Putonghua, report the language variety they were most commonly speaking, and give examples of their language use and forms of code-switching. Direct observation further allowed the researchers to evaluate the level of maintenance of Beijing dialect. The demographic details showed that the sample contained two groups of people, an Old Beijing group whose parents already lived in the area, and a New Beijing group, whose parents arrived in the city only more recently. Respondents attitude toward Beijing dialect and Putonghua was in agreement with earlier studies, comparing dialect and standard language (Fishman 1972). Beijing dialect was generally described as kind, pleasant to hear, giving a sense of belonging, creating a feeling of interest and of history, giving a sense of identity, a language one is accustomed to, and one that is efficient. The listing is in the order of frequency of reporting. In contrast the respondents feel that Putonghua is convenient for between group

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

communication, is considered normal and formal, whereas Beijing dialect sounds more like country-side speech, whereas Putonghua is easy to handle and easy to communicate in, and gives a sense of superiority. Clearly these answers more strongly reflect, respectively, Old Beijing backgrounds and New Beijing backgrounds, as we will see. Another interesting result was that respondents were not particularly clear as to the kind of language variety they were speaking. Having a set of criteria to evaluate a respondent’s speech, the researchers were able to categorize the answers given. One of the striking results of this approach was the conclusion that a large number of respondents claiming to speak Putonghua, actually were speaking Beijing dialect (n=44). After correcting for this discrepancy, the analysis showed that attitude, maintenance, selection and code-switching were all under the influence of the top three Beijing dialect speakers. A positive attitude toward Beijing dialect was also found when the spouse was a local, Beijing dialect speaking, person. That positive attitude was, further, more strongly supported in the five inner-city districts (as compared to the inner-suburb districts). The results for the Putonghua speakers was similar. Attitude toward Putonghua, selection of that variety instead of Beijing dialect, and code-switching were all three strongly supported when the first three listed speakers were Putonghua speakers. The second strong support came from having a New Beijing background. That background resulted in having a positive attitude toward Putonghua and having a preference for using it. A local spouse, further, not only helped to create a positive attitude but was also reported as instrumental in using Putonghua. In the same way, living in an inner-suburb district resulted not so much in a positive attitude toward Putonghua, but rather was supportive for selecting Putonghua as the preferred language variety. Is the triglossia hypothesis supported by these data? It seems difficult to make that claim on the basis of these data. Rather, these data add various dimensions to the triglossia/diglossia debate and allow the construction of a dialect maintenance model, that combines geographical, personal and network variables: a person’s background (Old Beijing versus New Beijing), the language preferences of the three persons he most frequently talks to, his place of residence (Inner district; Outer district), and the language preference and background of the spouse. This way a model is created that helps to understand the changing fortunes of dialect and standard language in a person’s personal life and can be applied to all dialect-standard language settings. Clearly, the potential difference between the established and newcomers is demonstrated in this study as well (Elias & Scotson 1965/1994). The model does not predict language behavior in public settings, such as public transport and shopping environments. Giving the overwhelming number of migrants in Beijing, the situation, most likely, will be similar to that reported for Sanya, increased use of Putonghua in public places. There of course one has no choice as to the selection of the kind of person one is interacting with, and the presence of a shared lingua franca will facilitate the choice of that language variety.

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

7. Conclusion We started the presentation with introduction of the concept long-term accommodation, based on Taiwanese data from the mid 1970s. When applying these to the Chinese mainland, we expected to find large scale adjustment toward the use of Putonghua, but for the first ten years after the ‘Open Door’ policy that could not be confirmed. Rather during those first ten years, Cantonese, the language of jobs and opportunity for a higher income, spread. A study of language acquisition of elementary school children in Hainan in more recent time, however, demonstrated that indeed the national standard language, Putonghua, the language of education, was spreading throughout the various groups making-up the Sanya community. The hypothesis based on that research stipulated that there would be nationwide a tendency to remove regional languages in favor of the national language Putonghua. A first test for the triglossia>diglossia hypothesis was provided by a real-time study of speech community formation. That study showed how, over-time, mutual adjustment between speakers takes place, and how, given the nature of a person’s occupation, place-of-origin, and social network, a speech community diversifies and gets socially stratified. This study for the first time demonstrated in detail how this mechanism works and how in the future we can understand language behavior in large-scale urban settings. Using these findings, we need to observe that the triglossia>diglossia hypothesis does not take into consideration the development of social stratification in Sanya, and that might very well force the authors to adjust the hypothesis in these terms. We expect dialect maintenance to be stronger among the lower educated working classes, whereas dialct loss will be stronger among elite groups.Finally, when testing this hypothesis in Shanghai and in Beijing, that hypothesis could not be supported. Additional forces needed to be added. Social stratification was already mentioned. The second addition to the discussion is the application of the full force of language vitality theory, which distinguishes between language status, demographic strength, and institutional support. The latter addition made clear that Shanghainese, as regional lingua franca, cannot be compared to a more local language variety such as Hainanese, which on each of the three dimensions mentioned has lesser status. The Beijing network density study, we introduced, allowed, in addition, the formulation of three more variables that need to be taken into consideration when a language variety is evaluated. One needs to consider the location and role of the city center as center of dialect maintenance, a speaker’s personal background (time and location of residence), network density (number of dialect speakers in the network) and the tradition of the family, which includes the language background of parents and spouse.

REFERENCES

BLOOMFIELD, L. 1933. Language. New York: Holt.

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

BOURHIS, RICHARD Y., GILES, HOWARD AND ROSENTHAL, DOREEN. 1981. Notes on construction of a ‘subjective vitality questionnaire’ for ethnolinguistic groups. Journal of Multilingual and Multicultural Development 2, 145-155. BROSNAHAN, L.F. 1963. Some historical cases of language imposition. In, Language in Africa, John F. Spencer (ed.), pp. 7-24. Cambridge: Cambridge University Press. CHU, XIAOQUAN. 2001. Linguistic diversity in Shanghai. Journal of Asian Pacific Communication,11 (1): 17-24. ELIAS, N. AND SCOTSON, J.L. 1965/1994. The established and the outsiders. London: Frank Cass & Co./ Norbert Elias Stichting FISHMAN, J. 1972. Sociolinguistics. Rowley, MA: Newbury House. GUO, XI. 2004. Zhongguo Shehui Yuyanxue (Chinese Sociolinguistics). Hangzhou: Zhejiang University Press. LABOV, W. 1966/2006. The Social Stratification of English in the New York City. New York: Cambridge University Press. PATRICK, PETER L. 2002. The Speech Community.” In J.K. Chambers, Peter Trudgill & Natalie Schilling-Estes, eds. The Handbook of Language Variation & Change. 573-597. Oxford: Blackwell. SONG, JING AND ZHU, YINGMEI. 2016. Attitude, maintenance, and use of Beijing dialect and Putonghua among younger generation Beijing born residents. Journal of Asian Pacific Communication, 26/1: 32-55. TANG, ZHIXIANG. 2016. Language choice and language attitude among different social segments in Shenzhen. Journal of Asian Pacific Communication 26/1: 143-160. VAN DEN BERG, M.E. 1986. Language Planning and language Use in Taiwan. Taipei: Crane. VAN DEN BERG, M.E. 2005. Vitality, Identity, and Language Spread: The Case of Shanghainese. Journal of Chinese sociolinguistics 5/2, 225-235. VAN DEN BERG, M.E. 2009. Jingji Xiandaihua he Yimin: Yueyu, Putonghua Jingzheng Guanxi de Fenxi. (Economic Modernization and Migration: An Analysis of Cantonese-Putonghua Competition.) Nanjing University Applied Linguistics Seminar, May 18. VAN DEN BERG, M.E. 2010. Socio-Economic Stratification in the Guangzhou Speech Community: Language Behaviour in Shopping Areas of Yuexiu and Tianhe Districts. In, M.E. van den Berg and Daming Xu (Eds.), (pp. 236-268). VAN DEN BERG, M.E. 2016. Modernization and the restructuring of the Shanghai speech community. Journal of Asian Pacific Communication, 26/1: 112-142. VAN DEN BERG, M.E. & XU, DAMING (EDS). 2010. Industrialization and the Restructuring of Speech Communities in China and Europe. Newcastle upon Tyne: Cambridge Scholars Publishing. XU, DAMING. 2004.Yanyu Shequ Lilun (Speech Community Theory). Journal of Chinese Sociolinguistics 4. XU, DAMING. 2006. Yuyan de bianyixing yu yanyushequ de yizhixing. (Variability of

VAN DEN BERG: CHINA URBAN LANGUAGE SURVEY

language and uniformity of Speech Communities.) Yuyan Jiaoxue yu Yanjiu 5. XU, DAMING. 2010. The Development of the Baotou Speech Community: A Quantitative Study of Nasal Variation in Mandarin Chinese. In, M.E. van den Berg and Daming Xu (Eds.), (pp. 120-140). XU, DAMING. 2016. Speech community theory and the language/dialect debate. Journal of Asian pacific Communication 16/1: 8-31. ZHAN, BOHUI. 1993. “Putonghua ‘nanxia’ yu Yue fangyan ‘beishang’ ” 普通话‘ 南下’ 与粤方言 ‘ 北上’ [“(Southbound” Putonghua and “Northbound” Cantonese)]. Xueshu Yanjiu 学术研究 4, 67-72.